EU Product Liability Directive: Transposition Countdown Links AI Act Non-Compliance to Strict Liability
As the December 9, 2026 transposition deadline approaches, European Union member states are rapidly preparing to incorporate the revised Product Liability Directive (PLD) into national law. This transition has been fundamentally altered by the European Commission's formal withdrawal of the proposed AI Liability Directive (AILD) in 2025.
The withdrawal of the AILD cements a dual regulatory framework: the EU AI Act serves as the compliance-based safety standard, while the 2024 Product Liability Directive (PLD) acts as the primary strict civil liability regime for claims brought by injured persons.
Strict Liability for AI-Enabled Products and Software
The 2024 PLD explicitly confirms that AI systems, software, and goods equipped with AI (including updates) are "products" under the Directive. This means that individuals can seek compensation when a defective AI product causes death, bodily injury, property damage, or data loss. Liable economic operators include the manufacturer, component manufacturer, importer, authorized representative, or fulfillment service provider.
The Presumption of Defectiveness and Shifted Evidentiary Burdens
The 2024 PLD introduces significant changes to the evidentiary framework, making it much easier for claimants to establish liability in complex AI scenarios. Under Articles 9–11, national courts can order defendants to disclose relevant evidence once the claimant has shown that their claim is plausible.
Crucially, the 2024 PLD establishes rebuttable presumptions of defectiveness and causation that shift the burden of proof to the defendant. As Bird & Bird explains:
"New presumptions of defectiveness are triggered by non-compliance with AI Act requirements or other EU sectorial legislation, technical complexity or failure to comply with an order to disclose evidence... These rules effectively shift the burden of proof, requiring defendants to demonstrate that the product was not defective or that any defect did not cause the alleged damage." — Bird & Bird, "AI Liability in light of the new 2024 PLD: expanded liability, challenging defences, and new evidentiary burdens" (February 17, 2026)
Under this system, non-compliance with the EU AI Act's safety and transparency obligations directly feeds into product liability litigation. If a high-risk AI system is deployed without meeting the AI Act's mandatory standards (such as data quality, technical documentation, continuous risk-management systems, or human oversight), the system will be presumed defective.
Action Plan for Enterprises
To prepare for the December 9, 2026 deadline, enterprises must:
- Map and Classify AI Systems: Identify all software and AI components integrated into products, and classify them under the EU AI Act's risk tiers.
- Prioritize Explainability and Logging: Ensure AI systems are designed with high transparency and robust logging to provide accessible explanations in the event of an audit or disclosure order.
- Establish Standardized Governance: Maintain rigorous documentation of safety compliance to rebut any potential presumption of defectiveness under the PLD.
- Reassess Insurance Policies: Review product liability, cyber, and professional indemnity coverages to eliminate gaps related to autonomous software or AI failures.